Results 1  10
of
101
Accessing Nearby Copies of Replicated Objects in a Distributed Environment
"... Consider a set of shared objects in a distributed network, where several copies of each object may exist at any given time. To ensure both fast access to the objects as well as efficient utilization of network resources, it is desirable that each access request be satisfied by a copy "clos ..."
Abstract

Cited by 549 (8 self)
 Add to MetaCart
Consider a set of shared objects in a distributed network, where several copies of each object may exist at any given time. To ensure both fast access to the objects as well as efficient utilization of network resources, it is desirable that each access request be satisfied by a copy &quot;close &quot; to the requesting node. Unfortunately, it is not clear how to efficiently achieve this goal in a dynamic, distributed environment in which large numbers of objects are continuously being created, replicated, and destroyed. In this paper
Probabilistic Approximation of Metric Spaces and its Algorithmic Applications
 In 37th Annual Symposium on Foundations of Computer Science
, 1996
"... The goal of approximating metric spaces by more simple metric spaces has led to the notion of graph spanners [PU89, PS89] and to lowdistortion embeddings in lowdimensional spaces [LLR94], having many algorithmic applications. This paper provides a novel technique for the analysis of randomized ..."
Abstract

Cited by 351 (32 self)
 Add to MetaCart
(Show Context)
The goal of approximating metric spaces by more simple metric spaces has led to the notion of graph spanners [PU89, PS89] and to lowdistortion embeddings in lowdimensional spaces [LLR94], having many algorithmic applications. This paper provides a novel technique for the analysis of randomized algorithms for optimization problems on metric spaces, by relating the randomized performance ratio for any metric space to the randomized performance ratio for a set of "simple" metric spaces. We define a notion of a set of metric spaces that probabilisticallyapproximates another metric space. We prove that any metric space can be probabilisticallyapproximated by hierarchically wellseparated trees (HST) with a polylogarithmic distortion. These metric spaces are "simple" as being: (1) tree metrics. (2) natural for applying a divideandconquer algorithmic approach. The technique presented is of particular interest in the context of online computation. A large number of online al...
On Approximating Arbitrary Metrics by Tree Metrics
 In Proceedings of the 30th Annual ACM Symposium on Theory of Computing
, 1998
"... This paper is concerned with probabilistic approximation of metric spaces. In previous work we introduced the method of ecient approximation of metrics by more simple families of metrics in a probabilistic fashion. In particular we study probabilistic approximations of arbitrary metric spaces by \hi ..."
Abstract

Cited by 266 (14 self)
 Add to MetaCart
(Show Context)
This paper is concerned with probabilistic approximation of metric spaces. In previous work we introduced the method of ecient approximation of metrics by more simple families of metrics in a probabilistic fashion. In particular we study probabilistic approximations of arbitrary metric spaces by \hierarchically wellseparated tree" metric spaces. This has proved as a useful technique for simplifying the solutions to various problems.
Concurrent Online Tracking of Mobile Users
 J. ACM
, 1991
"... This paper deals with the problem of maintaining a distributed directory server, that enables us to keep track of mobile users in a distributed network in the presence of concurrent requests. The paper uses the graphtheoretic concept of regional matching for implementing efficient tracking mechanis ..."
Abstract

Cited by 231 (7 self)
 Add to MetaCart
This paper deals with the problem of maintaining a distributed directory server, that enables us to keep track of mobile users in a distributed network in the presence of concurrent requests. The paper uses the graphtheoretic concept of regional matching for implementing efficient tracking mechanisms. The communication overhead of our tracking mechanism is within a polylogarithmic factor of the lower bound. 1 Introduction Since the primary function of a communication network is to provide communication facilities between users and processes in the system, one of the key problems such a network faces is the need to be able to Department of Mathematics and Lab. for Computer Science, M.I.T., Cambridge, MA 02139, USA. Email: baruch@theory.lcs.mit.edu. Supported by Air Force Contract TNDGAFOSR860078, ARO contract DAAL0386K0171, NSF contract CCR8611442, DARPA contract N0001489J 1988, and a special grant from IBM. y Departmentof Applied Mathematicsand Computer Science, The Weizm...
An Adaptive Data Replication Algorithm
 ACM Transactions on Database Systems
, 1997
"... This paper addresses the performance of distributed database systems. Specifically, we present an algorithm for dynamic replication of an object in distributed systems. The algorithm is adaptive in the sense that it changes the replication scheme of the object (i.e. the set of processors at which th ..."
Abstract

Cited by 185 (0 self)
 Add to MetaCart
(Show Context)
This paper addresses the performance of distributed database systems. Specifically, we present an algorithm for dynamic replication of an object in distributed systems. The algorithm is adaptive in the sense that it changes the replication scheme of the object (i.e. the set of processors at which the object is replicated), as changes occur in the readwrite pattern of the object (i.e. the number of reads and writes issued by each processor). The algorithm continuously moves the replication scheme towards an optimal one. We show that the algorithm can be combined with the concurrency control and recovery mechanisms of a distributed database management system. The performance of the algorithm is analyzed theoretically and experimentally. On the way we provide a lower bound on the performance of any dynamic replication algorithm.
Competitive NonPreemptive Call Control
"... We deal with randomized competitive algorithms for nonpreemptive call control on treelike switching networks. We give an optimal O(log n) competitive algorithm for nonpreemptive call scheduling on trees. We then extend the problem to include variable call rates, call durations, and arbitrary call ..."
Abstract

Cited by 105 (8 self)
 Add to MetaCart
We deal with randomized competitive algorithms for nonpreemptive call control on treelike switching networks. We give an optimal O(log n) competitive algorithm for nonpreemptive call scheduling on trees. We then extend the problem to include variable call rates, call durations, and arbitrary call benefits, and obtain a polylog competitive algorithm. We also show that many similar algorithms for different problems that can deal with constant values of parameters such as rates and benefits can be transformed into randomized algorithms that can deal with varying values of the parameters.
Data Replication for Mobile Computers
, 1994
"... Users of mobile computers will soon have online access to a large number of databases via wireless networks. Because of limited bandwidth, wireless communication is more expensive than wire communication. In this paper we present and analyze various static and dynamic data allocation methods. The ob ..."
Abstract

Cited by 89 (5 self)
 Add to MetaCart
(Show Context)
Users of mobile computers will soon have online access to a large number of databases via wireless networks. Because of limited bandwidth, wireless communication is more expensive than wire communication. In this paper we present and analyze various static and dynamic data allocation methods. The objective is to optimize the communication cost between a mobile computer and the stationary computer that stores the online database. Analysis is performed in two cost models. One is connection (or time) based, as in cellular telephones, where the user is charged per minute of connection. The other is message based, as in packet radio networks, where the user is charged per message. Our analysis addresses both, the average case and the worst case for determining the best allocation method. 0 1 Introduction Users of mobile computers, such as palmtops, notebook computers and personal communication systems, will soon have online access to a large number of databases via wireless networks. The ...
Approximation Algorithms for Data Placement in Arbitrary Networks
, 2001
"... We study approximation algorithms for placing replicated data in arbitrary networks. Consider a network of nodes with individual storage capacities and a metric communication cost function, in which each node periodically issues a request for an object drawn from a collection of uniformlength objec ..."
Abstract

Cited by 84 (4 self)
 Add to MetaCart
(Show Context)
We study approximation algorithms for placing replicated data in arbitrary networks. Consider a network of nodes with individual storage capacities and a metric communication cost function, in which each node periodically issues a request for an object drawn from a collection of uniformlength objects. We consider the problem of placing copies of the objects among the nodes such that the average access cost is minimized. Our main result is a polynomialtime constantfactor approximation algorithm for this placement problem. Our algorithm is based on a careful rounding of a linear programming relaxation of the problem. We also show that the data placement problem is MAXSNPhard. We extend our approximation result to a generalization of the data placement problem that models additional costs such as the cost of realizing the placement. We also show that when object lengths are nonuniform, a constantfactor approximation is achievable if the capacity at each node in the approximate solution is allowed to exceed that in the optimal solution by the length of the largest object.
Data Partitioning and Load Balancing in Parallel Disk Systems
, 1994
"... Parallel disk systems provide opportunities for exploiting I/O parallelism in two possible ways, namely via interrequest and intrarequest parallelism. In this paper we discuss the main issues in performance tuning of such systems, namely striping and load balancing, and show their relationship to ..."
Abstract

Cited by 81 (8 self)
 Add to MetaCart
(Show Context)
Parallel disk systems provide opportunities for exploiting I/O parallelism in two possible ways, namely via interrequest and intrarequest parallelism. In this paper we discuss the main issues in performance tuning of such systems, namely striping and load balancing, and show their relationship to response time and throughput. We outline the main components of an intelligent file system that optimizes striping by taking into account the requirements of the applications, and performs load balancing by judicious file allocation and dynamic redistributions of the data when access patterns change. Our system uses simple but effective heuristics that incur only little overhead. We present performance experiments based on synthetic workloads and reallife traces.